If you’re good at something, you usually get rewarded with more work. This apparently applies to AI as well. The latest trends and directions of AI development is pointing towards AaaS: Agent as a Service, where AI solutions are going to not only find the right answers for you but also actually conduct some, or all of, the work that needs to be done using that answer. If you’ve ever dreamt of having your own Jarvis from Iron Man, the future is actively marching closer to be your present. Or is it? 

This post will share some of my latest learning around the topic of AaaS with some thoughts on what you could be doing to not fall behind the curve on the topics of AI. 

Let’s start with defining what AaaS is and why it may be of any importance for you. Many AI solutions that are out there today are extremely powerful and versatile but they mostly reside at a transactional level of interaction with the human users. An agentic solution would be able to take a step further and complete a certain task or set of tasks that it ‘knows’. 

For instance, I use ChatGPT for all sorts of tasks. If I were writing a blog post with it, I’ll be able to ask questions for ideating and even ask it to write a draft that’s 2 pages long. But I’ll have to download this draft, read through it, make edits myself or ask ChatGPT to do it, then upload it onto the WordPress site. I can leverage ChatGPT at any or all of the steps but the interaction is transactional: I ask it to do something, and it does it for me. 

An agentic AI solution would be able to not only ideate on a topic and write a draft, but also autonomously review it and suggest edits, finalize an acceptable version of the copy and email it to my while simultaneously uploading it to the social media accounts that it has access to, all without my direct intervention. If this sounds a bit too SciFi-ish, it’s not. 

The velocity on this development is such that you could now use a few tools yourself like the OpenAI Platform, n8n, Lovable, yourself like to make an AI agent ‘system’ right now. I’m actually in the process of learning some of these tools and I’ll plan on sharing my journey in a once it’s at a better spot. 

You might be thinking, 

“All of this is cool but why should I care? My life is busy as it is!”

Well, I think it’s important because of 2 reasons: You can make your life a little less busy or you might end up with a life that’s less busy. There was a lot of rumbling when generative AI was first introduced where folks worried that they may losing their jobs. These were valid to some degree but I thought the hype around AI was a tad bit overrated. But this is the first time where I actually think the full power of this technology, once adopted and tested, could lead to a potential decreases in the human workforce. Like any technological disruption, those of who can harness its benefits will be able to enjoy a significant advantage above those who can’t, and with how quickly things are changing with AI, it’ll soon become nearly impossible to follow the trends unless you’ve had a head start. 

The key difference that triggered a change in my perception of job security is the incorporation of “action” into the AaaS concept. I always thought the ability to act on behalf of the human user would be what’s needed to blow things out of proportion and it seems that future isn’t that far away. But how far away is it?

So how far are we exactly from witnessing some major disruptions? Nobody can be sure of the timing, but everyone could agree that it’s within a year or two until we see some big tide changes. 

The concept of AaaS is still in its infancy (despite its steroid fueled growth) and as such, there are a limited number of software solutions, websites, and other platforms that can be connected by an agentic API. This creates a bottleneck to expanding the list of actions that can be taken by the agent. However, I wouldn’t be surprised one bit if every major SaaS provider out there today has been aware of the changing directions and already have competitors or ongoing initiatives to enable their own agentic solutions by the end of 2025. 

Here are some examples of pioneering pathfinders in this arena and what they’re working on. 

Vertical agents are specialized solutions that are fine tuned to the tasks required in a specific field. For instance Zora AI is adopted by Deloitte to assist with financial audits, lead generation, nurturing, and CRM management. Agent CRM is another example where it manages marketing, lead nurturing, and sales management. Vertical agents are generally much faster to create, deploy, and scale due to its defined scope and specificity of the tasks. It sacrifices versatility as you might imagine from platforms like ChatGPT but it is accurate and effective for the area that it’s meant to support. 

Compared to vertical agents, horizontal agents will be something that’s applicable for multiple industries and hence needs to be flexible while retaining all the efficacies and efficiencies of an AI agent. With the integration with OpenAI, you can already start building agents with Microsoft Copilot. You can make agents that manage your emails, calendars, projects, and leads. Check out their website here!

Ok so agentic AI solutions are on the rise and things are expected to change even more rapidly. Now what? What does this mean for you and what should you be doing now? Largely, I think you have 3 options.

Being able to effectively search, adopt, and leverage agentic solutions will be a crucial skill set that needs to be developed for everybody that wants to stay relevant. A versatile transactional solution like ChatGPT can significantly reduce your workload depending on how you use it. With effective agentic solutions available, the difference in productivity associated with those that can use these solutions versus those who don’t will be as dramatic as Tony Stark with Jarvis versus Peter Parker with his school laptop. To stay relevant, stay on top of the new solutions that will continuously be introduced and follow the latest trends on how those solutions are used. You might also want to shift your perception on AI from ‘using the best solution’ to ‘assembling the right team of solutions’, which will become your set of AI agents. 

What really blew me away about the latest trends around agentic AI is how little coding it requires. If you’re leveraging the right tools and their offerings, you can virtually get by without doing any coding at all. Instead, it’ll be a simple drag and drop process to build your workflows and a few copying and pasting of credentials to link your agent to the platforms of your desire. You might need to brush up on a few concepts about logic building but it really shouldn’t be anything too complicated for anybody that can write an if-then statement. 

Let’s be real. You saw this one coming. To really stand out in the race of agentic AI, you really need to be both. Agentic AI strategies can adopt multiple AI components with specialized tasks to deliver the best performance. If you’re an active user, you might want to be able to utilize the right solutions at the right time. But to do so, you’ll probably have to understand the ‘behind-the-curtains’ details enough to be able to customize your own agents and tailor it to your needs. 

Eventually, I have no doubt that this will become as easy as using the Microsoft Office suite but until then, follow this blog to stay on top of the latest AI news and also share your learnings around using and implementing agentic AI!

If you’re looking for any specific content or have any questions on this post, leave a comment below or contact me directly anytime!